Notes on draft version 2.3

The most important change:

The proton rescattering treatment is completely updated.  The website 
links to a note with more details, plots, and tables.  As expected, the 
changes to MA are minor, though other things in the paper change 
also.  This, and the two-track to 1-track migration are discussed more 
completely in this draft than they were before.

The changes in the proton rescattering treatment will require that I 
update several of the numbers and figures in the paper (some but not all 
are completed), but that should not detract from your review of this 
draft.

If you choose to compare numbers with the previous draft, here is a quick 
summary of what changed in the procedure and in the results:

1.  Take the proton rescattering error to be 1.00 +/- 0.20 (not 0.87 +/- 0.10).
2.  Use the new reweigting parameterization that I produced from Neut45.

Item 1. has the more significant effect, shifting MA by 0.04, but also the 
nonQE/QE, and the 2tk/1tk migration fit values a fair amount.  I have also 
included proton rescattering in the table of parameters that have an 
effect, now estimated at 3% because of the increased +/- 0.2 on the 
parameter.  Item 2. has very little effect, as expected.

Other changes:

The delta-theta plot is changed (no longer cos(delta-theta), following the 
suggestion.  I also like how it communicates the separation.  I have not, 
and I do not plan to include yet another line describing proton 
rescattering effects in this plot.  The numtracks plot is updated to fix a 
calculation bug, but the zero proton rescattering line is still not 
directly from a SciFi MC sample, it is the same 20% sample Chris used in 
NuInt01. However, this 20% figure is approximately correct, I have 
estimated it from the 80% 100% proton rescattering comparison.  I 
personally still have doubts that this *plot* should be included, though I 
completely agree that this 20% should be stated clearly in the text.

In addition to the changes in Proton Rescattering, I found an error in how 
the chisquare is accounted for in the 1tk to 2tk migration parameters.  
This had essentially no effect in anything but the chisquare and dof 
accounting, and is now fixed.

Two figures (of SciFi detector) are temporarily removed due to printing 
problems.  Chris originally reported this problem, and I have reproduced 
it, but not solved it here.  I will solve these before a draft circulates 
to the collaboration.

The text has been proofread and updated in many places, following your 
suggestions.